INDEX

Explanations

returning values or indicating function end

This neuron fires on return statements (the keyword “return”).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

the

-1.92

 Their

-1.66

 their

-1.41

 With

-1.39

 nationalen

-1.38

 these

-1.38

を取る

-1.27

</h5>

-1.26

⿸

-1.25

🔥🔥

-1.24

POSITIVE LOGITS

.}(

1.82

 will

1.50

iramente

1.47

’

1.47

MAXN

1.45

了一眼

1.42

佉

1.38

یزات

1.36

issipp

1.35

Activations Density 0.065%